A tip - use print(df.columns) before trying df.usecols. This shows the headers, more importantly it shows any leading spaces which may not be obvious.